Semantic Textual Similarity On Sts12

Métriques

Spearman Correlation

Résultats

Résultats de performance de divers modèles sur ce benchmark

Nom du modèle	Spearman Correlation	Paper Title	Repository
PromptEOL+CSE+LLaMA-30B	0.7972	Scaling Sentence Embeddings with Large Language Models	-
PromptEOL+CSE+OPT-13B	0.8020	Scaling Sentence Embeddings with Large Language Models	-
Dino (STSb/̄	0.7027	Generating Datasets with Pretrained Language Models	-
SRoBERTa-NLI-large	0.7453	Sentence-BERT: Sentence Embeddings using Siamese BERT-Networks	-
Trans-Encoder-BERT-large-bi (unsup.)	0.7819	Trans-Encoder: Unsupervised sentence-pair modelling through self- and mutual-distillations	-
Trans-Encoder-RoBERTa-large-cross (unsup.)	0.7828	Trans-Encoder: Unsupervised sentence-pair modelling through self- and mutual-distillations	-
SimCSE-RoBERTa-base	0.7016	SimCSE: Simple Contrastive Learning of Sentence Embeddings	-
AnglE-LLaMA-7B	0.7868	AnglE-optimized Text Embeddings	-
Mirror-BERT-base (unsup.)	0.674	Fast, Effective, and Self-Supervised: Transforming Masked Language Models into Universal Lexical and Sentence Encoders	-
DiffCSE-RoBERTa-base	0.7005	DiffCSE: Difference-based Contrastive Learning for Sentence Embeddings	-
PromptEOL+CSE+OPT-2.7B	0.7949	Scaling Sentence Embeddings with Large Language Models	-
PromCSE-RoBERTa-large (0.355B)	0.7956	Improved Universal Sentence Embeddings with Prompt-based Contrastive Learning and Energy-based Learning	-
DiffCSE-BERT-base	0.7228	DiffCSE: Difference-based Contrastive Learning for Sentence Embeddings	-
AnglE-LLaMA-13B	0.7868	AnglE-optimized Text Embeddings	-
Trans-Encoder-RoBERTa-base-cross (unsup.)	0.7637	Trans-Encoder: Unsupervised sentence-pair modelling through self- and mutual-distillations	-
Trans-Encoder-BERT-base-bi (unsup.)	0.7509	Trans-Encoder: Unsupervised sentence-pair modelling through self- and mutual-distillations	-
SimCSE-RoBERTa-large	0.7746	SimCSE: Simple Contrastive Learning of Sentence Embeddings	-
Mirror-RoBERTa-base (unsup.)	0.648	Fast, Effective, and Self-Supervised: Transforming Masked Language Models into Universal Lexical and Sentence Encoders	-
IS-BERT-NLI	0.5677	An Unsupervised Sentence Embedding Method by Mutual Information Maximization	-
BERTlarge-flow (target)	0.6520	On the Sentence Embeddings from Pre-trained Language Models	-

0 of 20 row(s) selected.